Full length articleBEM-based iterative EKF channel estimation using superimposed pilot for OFDM system in high-speed railway
Introduction
The downlink channel estimation is an important part of the receiver, which is designed for orthogonal frequency division multiplexing (OFDM) system. Traditional channel estimation methods often use pilot symbol assisted modulation (PSAM) estimation method [1], [2], [3]. The method inserts pilots between data symbols by means of time division multiplexing (TDM) in the stage of estimation. The channel state information (CSI) of the entire transmission channel can be calculated by the position of pilots. In the time-invariant or slow time-varying channel, since the channel information during an OFDM symbol remains substantially stable and the TDM pilot is the preferred pilot sequence, the estimation performance of the PASM channel estimation method is better. Bandwidth loss maybe ignored when the frequency insertion period is large enough. However, in the high-speed mobile environment, since the wireless channel is affected by the multipath spread and the Doppler effect, the channel changes significantly within one OFDM symbol and the channel has double-selective fading and time-domain non-stationary characteristics [4], [5], [6], which means that the inter carrier interference (ICI) will seriously affect the performance of the traditional channel estimation method. Therefore, the TDM pilot must be inserted frequently to track the channel change if the traditional estimation method is still used. However, the insertion of the pilot sequence will occupy the time slot or subcarrier and the system frequency band will be wasted, resulting in a severe drop in transmission rate and spectrum efficiency. In order to cope with the double-selective fading and time-domain non-stationary characteristics of the channel in high-speed mobile environment and the low spectrum utilization of the system using the traditional channel estimation method, it is necessary to find a channel estimation method with better estimation performance.
Since the time-domain channel estimation method can directly estimate the channel impulse response (CIR) on each path, it can be widely used to eliminate inter subcarrier interference (ICI) effectively in the channel estimation [7]. However, under the double-selective fading channel, the time-domain channel estimation method requires more channel estimation parameters. It is hard to get the CIR, while the basis expansion model (BEM) has better compression performance. The BEM is approximated to model the OFDM channel coefficients, which can reduce the amount of parameters estimated and the auxiliary time-domain channel estimation method eliminates the serious ICI interference caused by the double-selective fading. For the non-stationary characteristics of the channel, according to our previous research results [8] [9], the Kalman filtering (KF) method can better track the time-varying time-domain channel.
On the other hand, in order to solve the problem of low spectrum utilization caused by the traditional channel estimation method in high-speed railway scenarios, the channel estimation method using superimposed pilot has received extensive attention [10], [11], [12]. Nevertheless, the interference from data symbols has greatly deteriorated the performance of channel estimation and data detection in the superimposed pilot scheme [11]. Therefore, researchers have done a lot of exploration to solve the problems. [10] proposes a channel estimation method for superimposed pilot sequences (ST) for wireless communication, the method results in a higher peak-to-average ratio (PARR) and the estimated performance is severely degraded due to interference from unknown transmitted data signals. In [12], the channel estimation method of orthogonal superimposed pilot (OSP) is proposed, which sacrifices a small amount of transmission rate to ensure that the pilot and data at the receiver are decoupled from each other, making the channel estimation process more complicated. It can be seen that the decoupling of pilot and data information is the key to improve the performance of such channel estimation [13]. In addition, in order to estimate the CSI accurately, the channel estimation method of superimposed pilot is different from the PASM estimation method. As the estimation performance requirement increases, it is necessary to increase the pilot allocation power, which is undoubtedly a big challenge to obtain accurate CSI when power is limited [14], [15]. As a result, although there are still some technical challenges in channel estimation based on superimposed pilot, the advantages of not losing the spectrum efficiency of the system and improving the overall performance of the wireless communication system are very prominent. Hence, in order to improve the spectrum efficiency of the system and communication performance in the high-speed mobile communication scenario, this paper focuses on the channel estimation method using superimposed pilot of OFDM system in high-speed railway with double-selective fading and time-domain non-stationary channel characteristics.
The main contributions of this paper are as follows. Firstly, the channel estimation method of EKF is proposed, and the channel is modeled as a joint estimation of CIR and time-varying time correlation coefficients. Secondly, to eliminate the data symbol interference at the superimposed pilot position, the pilot and data at the superimposed position are decoupled, and the transmitted symbols are reconstructed. Then the reconstructed symbol is used to update the weighted matrix in the EKF observation equation and improve estimation accuracy by iteration.
The organization of this paper is as follows. In Section 2, a system model using superimposed pilots is introduced, and the time-domain channel is modeled as an auto regressive (AR) process. In Section 3, an iterative estimation scheme for EKF is proposed, including initial channel estimation, decoupling and reconstruction of pilot and data, and design of iterative estimation. In Section 4, the complexity of the proposed algorithm and the similar channel estimation algorithm are summarized and analyzed. In Section 5, the estimated performance of the proposed method and the traditional channel estimation method at different velocities are compared on the MATLAB simulation platform. Finally, the conclusions are drawn in Section 6.
Section snippets
System model
In order to solve the problem of low spectrum utilization caused by the channel estimation in high-speed environment, it is important to choose a suitable pilot structure. Indeed, the comb pilot is to insert pilot symbols in the frequency domain at a certain frequency interval, and the other subcarrier except the pilot symbols transmit data symbols and this pilot structure is suitable for channels with faster channel changes. However, block pilots are used in this paper where all subcarriers
Channel estimation method
Since the CIR has time-varying characteristic in high-speed environment, the channel estimation methods of conventional superimposed pilot cannot track the channel well. Therefore, the BEM-based iterative EKF (BEM-iEKF) channel estimation method is innovatively proposed in this section. Firstly, the CIR and time domain correlation coefficient are jointly estimated to track the channel variation in real time. Then in order to improve estimation accuracy and eliminate the interference between
Computational complexity analysis
Table 1 compares the computational complexity of several classical channel estimation methods and the same channel estimation method in completing one OFDM symbol time (the number of times the algorithm needs to perform once).
In Table 1, since no prior knowledge is needed, the ST channel estimation method [10] based on traditional superposition training sequence only requires simple weighted averaging and pseudo-inverse operations, so the complexity is low. However, the data superimposed on the
Simulation analysis
This paper uses MATLAB to simulate and analyze the proposed SP-BEM-iEKF channel estimation method and other similar channel estimation methods and classical channel estimation methods. The parameters of simulation system are shown in Table 2.
The accuracy of the estimation and the performance of the receiver in different velocity scenario of ST channel estimation method, the traditional pilot assisted EKF channel estimation method, OSP based channel estimation method as well as SP-BEM-iEKF
Conclusion
In this paper, for the weakness of the traditional channel estimation methods in high-speed railway, we propose a BEM-iEKF channel estimation method using superimposed pilot. The simulation results show that, compared with traditional channel estimation methods, the proposed method has a better accuracy of estimation and the whole performance of the system with the increase of the number of iterations, and the system has a higher effective data throughput, which is more suitable for high-speed
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (No. 61501066), the Natural Science Foundation of Chongqing, China (No. cstc2019jcyj-msxmX0017).
Yong Liao received the Ph.D. degree from Chongqing University, Chongqing, China, in 2014. He is a Research Associate and the Deputy Director with the Key Laboratory of Aircraft TT&C and Communication, Ministry of Education, Chongqing University, Chongqing, China. His research interests include high-speed mobile communication, 5G and future communication, and aerocraft TT&C and communication. Email: [email protected]
References (23)
- Jiang Xuehua, Chen Peijiang, Study and implementation of MIMO-OFDM system based on matlab, in: IEEE International...
- et al.
Wireless Communications over Rapidly Time-Varying Channels
(2011) - et al.
An adaptive geometry-based stochastic model for nonisotropic MIMO mobile-to-mobile channels
IEEE Trans. Wireless Commun.
(2009) - et al.
Challenges toward wireless communications for high-speed railway
IEEE Trans. Intell. Transp. Syst.
(2014) - Pei-Yun Tsai, Tzi-Dar Chiueh, Frequency-domain interpolation-based channel estimation in pilot-aided OFDM systems, in:...
- et al.
Pilot-assisted time-varying channel estimation for OFDM systems
IEEE Trans. Signal Process.
(2007) - et al.
A non-stationary IMT-A MIMO channel model for high mobility wireless communication systems
IEEE Trans. Wireless Commun.
(2017) - et al.
EKF-Based joint channel estimation and decoding design for non-stationary OFDM channel
- et al.
Non-stationary and doubly-selective channel estimation method based on basis expansion model
Acta Electronica Sinica.
(2019) - et al.
Orthogonal superimposed training design for doubly selective channel estimation using basis expansion models
Trans. Emerg. Tel Tech.
(2017)
Cited by (3)
EM-EKF fast time-varying channel estimation based on superimposed pilot for high mobility OFDM systems
2021, Physical CommunicationCitation Excerpt :To sum up, the process of EM-EKF fast time-varying channel estimation algorithm based on superimposed pilot is shown in Algorithm 1. This paper mainly compares the performance of the iterative BEM-EKF algorithm [11], the EM algorithm [20] based on superimposed pilot and the EM-EKF algorithm proposed in this paper in two indexes of normalized mean square error and bit error rate. Fig. 2 shows the NMSE performance under different iterations when SNR=20 dB.
Estimation of bit error rate in 2×2 and 4×4 multi-input multioutput-orthogonal frequency division multiplexing systems
2023, International Journal of Electrical and Computer Engineering
Yong Liao received the Ph.D. degree from Chongqing University, Chongqing, China, in 2014. He is a Research Associate and the Deputy Director with the Key Laboratory of Aircraft TT&C and Communication, Ministry of Education, Chongqing University, Chongqing, China. His research interests include high-speed mobile communication, 5G and future communication, and aerocraft TT&C and communication. Email: [email protected]
Nan Zhang received B.S. degree from China West Normal University, Nanchong, China, 2017, and is currently studying for M.S. degree of Chongqing University, Chongqing, China. Her current research direction is wireless communications channel estimation in high mobility scenarios. Email: [email protected]
Guodong Sun received B.S. degree from Chongqing University, China, 2018, and is currently studying for M.S. degree of University, Chongqing, China. His current research direction communications channel estimation and channel interpolation algorithms. Email: [email protected]
Xinyi Yang received B. S. degree from Chongqing University of China, Chongqing, China, 2018, and is currently studying for M.S. degree of Chongqing University, Chongqing, China. Her current research direction is precoding algorithm in high-speed mobile communication. Email: [email protected]
Ning Sun received B. S. degree from Hebei University of China, Baoding, China, 2018, and is currently studying for M.S. degree of Chongqing University, Chongqing, China. His current research direction is deep learning in wireless communication applications. Email: [email protected]
Yuanxiao Hua received B. S. degree from North University of China, Taiyuan, China, 2016, and is currently studying for M.S. degree of Chongqing University, Chongqing, China. His current research direction is wireless communications channel estimation and deep learning in communication applications. Email: [email protected]
Haimei Yao received the B.S. degree from Jiangxi University of Science and Technology, Ganzhou, China, 2016, and is currently studying for M.S. degree of Chongqing University, Chongqing, China. Her current research direction is AI and its Application in Wireless Communication. Email: [email protected]